Struct AmxBf16 

pub struct AmxBf16;

x86/x86_64 AMX BF16 matrix multiply backend.

Trait Implementations

impl AmxGemm<Bf16, Bf16, F32> for AmxBf16

unsafe fn amx_gemm( m: usize, n: usize, k: usize, a: *const Bf16, a_stride: usize, b: *const Bf16, b_stride: usize, c: *mut F32, c_stride: usize, )

Perform matrix multiplication c += a * b using Intel AMX block execution.

impl Clone for AmxBf16

fn clone(&self) -> AmxBf16

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for AmxBf16

impl Debug for AmxBf16

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl Eq for AmxBf16

impl PartialEq for AmxBf16

fn eq(&self, other: &AmxBf16) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl SimdArch for AmxBf16

const NAME: &'static str = "amx_bf16"

Human-readable ISA name ("avx2", "avx512", "neon", "scalar").

const REGISTER_WIDTH_BITS: u32 = 8192

Width of vector registers in bits.

const ISA_FAMILY: IsaFamily = IsaFamily::X86

ISA family for this architecture.

const FMA_THROUGHPUT_HINT: u32 = 16

Suggested TILE_M value for tiled_dot to saturate FMA throughput.

fn is_runtime_supported() -> bool

Returns true when the current host may execute this architecture’s native instructions from safe wrappers.

impl StructuralPartialEq for AmxBf16

impl TileMatrixMultiply<Bf16, Bf16, F32, AmxBf16, AmxBf16, 16, 16, 32> for AmxBf16

unsafe fn tile_matmul( c: *mut F32, c_stride: usize, a: *const Bf16, a_stride: usize, b: *const Bf16, b_stride: usize, )

Performs tile matrix multiplication: C += A * B